Short answer
Decision Intelligence is the discipline of structuring important choices by connecting evidence, causal understanding, alternatives and human judgement to the outcomes leaders are trying to achieve.
In healthcare, that means starting with the outcome and the decision, then bringing together data, science, AI, clinical expertise, operational knowledge and lived experience to understand what should change and what is likely to happen next.
It is not another name for analytics. Analytics informs the decision. Decision Intelligence structures the choice, its consequences and the action that follows.
Why healthcare decisions are different
Healthcare organisations are organised by functions, but important decisions cut across them.
Finance, workforce, operations, clinical quality and population health often hold different pieces of the same problem. Each function sees the organisation through its own data, accountabilities and professional language.
Consider a trust with a fragile clinical service. Finance sees rising expenditure. Workforce analysis shows rota gaps. Operational data shows cancellations and transfers. Quality indicators may not yet have crossed a regulatory threshold. Clinicians disagree about whether the service needs investment, a network arrangement or consolidation.
None of those perspectives is sufficient on its own. More importantly, the service is not experiencing five separate problems. It is experiencing one problem viewed from five functions.
Conventional management processes tend to preserve those divisions. Separate teams produce separate reviews and take separate papers through separate committees. The board eventually has to reconcile them, often while the decision itself remains loosely framed.
Decision Intelligence takes the opposite approach.
Decision Intelligence reconstructs the whole decision from evidence that organisations usually hold in separate functions.
This matters because healthcare choices create consequences beyond the function or institution making them. Reducing beds changes flow, workforce pressure and community demand. Consolidating a service may improve clinical resilience while increasing travel time. Moving care out of hospital can improve access but weaken a provider’s income if money and responsibility do not move with the activity.
The unit of analysis must therefore be the whole decision, not the report produced by one part of the organisation.
Reconstructing the whole decision
A decision is more than a problem statement or ambition.
“Improve productivity” is not a decision. Neither is “address the workforce challenge” or “move care closer to home”. A decision requires a choice between credible courses of action.
For example:
- Should this clinical service remain standalone, operate through a network or consolidate?
- Should available investment expand hospital capacity or strengthen an earlier part of the pathway?
- Should a trust recruit more consultants or redesign how the existing team works?
- Which financial opportunities are genuinely recoverable, and which remain theoretical?
Once the decision is explicit, leaders can determine what evidence bears on it. A clinical-service choice may require demand, outcomes, workforce, finance, dependencies, patient access and alternative models of care. Some of that evidence will be quantitative. Some will come from professional experience or the lived experience of patients and staff.
Decision Intelligence then traces how each option is expected to affect the system. What action follows from the choice? Which relationships or behaviours need to change? Where will benefits appear? Who carries the risk? What assumptions must hold?
From outcome to learning
An explanatory sequence, not a proprietary method. The formal Strasys method is the SMASH approach below.
Both feed causal understanding. Neither makes the decision. Step 04 stays with the people and organisations authorised to decide, and step 07 returns to step 01.
The process does not end when a board approves an option. Responsibility, resources and measures must follow. Results then need to be compared with the original assumptions. That is the learning loop: not another evaluation exercise, but a way to improve the next decision.
AI, expertise and judgement
AI is a component of Decision Intelligence, not its definition.
AI extends analysis. It can process large datasets, identify patterns, generate forecasts and simulate possible consequences. It can also reproduce poor assumptions or hidden biases at scale. Its contribution depends on the evidence, model and question it is given.
Expertise interprets context. Clinical, operational and financial practitioners understand constraints, dependencies and behaviours that are absent from formal datasets. Patients, staff and communities contribute lived experience of how the system actually works.
Judgement makes the accountable decision. Leaders weigh evidence, uncertainty, values, risk and competing interests. Some consequences cannot sensibly be collapsed into one score. A board may have to balance resilience against access, or system value against the financial position of one organisation.
Judgement is not a licence to ignore evidence. Nor does stronger analysis remove the need for judgement. Decision Intelligence makes the relationship between the two visible, which is also why we treat accountable human judgement as a governance question rather than a technical one.
The distinction is deliberate:
AI extends analysis. Expertise interprets context. Judgement makes the accountable decision.
Healthcare decisions affect clinical risk, public access, staff livelihoods and the distribution of resources. Accountability must remain with the people and organisations authorised to decide.
Start with the outcome
Most organisations begin with the data they already collect. They ask what the dashboard shows, which measures are deteriorating and where performance differs from plan.
Decision Intelligence starts earlier. It asks what outcome is sought and what decision has to be made to influence it.
This prevents analytical effort from being shaped by the available data alone. A dataset may measure activity well while saying little about whether the activity was needed or whether it improved people’s lives.
Clayton Christensen’s Jobs to Be Done theory offers a useful supporting lens.5 It asks what progress a person is trying to make in particular circumstances rather than starting with the product or service being supplied.
In healthcare, a person’s job may be to remain well enough to work, recover sufficiently to care for their family or manage a condition without organising their life around the NHS. The institutional task may be to deliver an appointment, test or procedure.
A service can complete its institutional task while failing the consumer’s job.
We use this lens to clarify the outcome sought before modelling service, workforce or system choices. It is one input to Decision Intelligence, not a definition of the discipline.
Lorien Pratt’s work provides another useful principle. She describes Decision Intelligence as a bridge between decisions and data and emphasises the causal path from actions to outcomes.3 This reinforces the need to ask how a proposed choice is expected to change the system, rather than treating correlation or prediction as sufficient.
External thinkers help explain these principles. We apply and extend them in healthcare through our own decision architecture, population-need approach and methods.
How DI differs from BI and AI
| Discipline | Primary job | What remains unresolved |
|---|---|---|
| Business intelligence | Describes what happened and where performance varies. | What leaders should choose and why. |
| Analytics | Explains patterns and estimates what may happen next. | Which outcome matters and which option should be preferred. |
| Artificial intelligence | Detects patterns, generates forecasts and supports defined tasks. | Accountable judgement about values, risk and trade-offs. |
| Decision Intelligence | Structures the path from outcome and choice through consequences, action and learning. | It still depends on good evidence, expertise and implementation. |
A sophisticated dashboard can leave a board debating several incompatible versions of the problem. A predictive model can forecast demand without answering whether the system should add capacity, redesign a pathway or intervene earlier.
Decision Intelligence gives BI, analytics and AI a specific role within a decision. Their purpose is to reduce uncertainty that matters to the choice.
How Strasys applies it
We apply Decision Intelligence to healthcare by starting with population and consumer need, reconstructing the full decision across clinical, workforce, operational and financial evidence, and tracing choices through to outcomes.
This is expressed through the SMASH approach and Decision Continuum.
Most healthcare planning begins with what the organisation already provides. It examines current services, workforce and budgets, then asks how they might improve.
SMASH reverses the direction. It starts with the consumer and works back through the system. What does the population need? What job is the person trying to get done? What outcome should the system produce? What service response, workforce, capacity, organisation and resource model would that require?
The SMASH Decision Continuum
Strasys works consumer to policy. Conventional planning works policy to patient.
Explore the interactive Decision Continuum, and the products mapped to each step, on the Platform page.
This shift matters because inherited services and organisational boundaries are no longer treated as fixed. The organisation becomes one part of the answer rather than the starting point.
We then bring together the evidence required to understand the choice. Different scientific and professional disciplines contribute where they are useful: systems thinking for relationships and dependencies; system dynamics for feedback and delay; behavioural and social science for human response; economics for incentives and allocation; mathematics, data science and AI for patterns, uncertainty and scenarios; clinical and operational expertise for safety, context and feasibility.
The disciplines do not disappear. They work on the same decision.
This approach connects our four authority territories:
- In financial sustainability, it distinguishes theoretical opportunities from trapped value in healthcare that is accessible, recoverable and realised, which the Strasys Value Index is built to locate.
- In clinical-service sustainability, it connects need, workforce, quality, economics, dependencies and alternatives.
- In workforce design, it links service demand to roles, deployment, capacity, behaviour and cost.
- In health-system design, it starts with population need and works back through services, workforce, organisations, finance and policy.
Decision Intelligence is the intellectual framework connecting these areas. It is not a separate body of abstract theory detached from the decisions healthcare leaders face.
Evidence and limitations
Established external evidence
Gartner defines Decision Intelligence as a practical discipline that advances decision making by explicitly understanding and engineering how decisions are made, and how outcomes are evaluated, managed and improved through feedback.1 That framing supports the general definition of the field, although it is not specific to healthcare.
The World Health Organization defines evidence-informed decision-making as identifying, appraising and mobilising the best available evidence for health policy and programmes.2 Decision Intelligence builds on that discipline by connecting evidence to alternatives, consequences, judgement, action and learning.
Lorien Pratt’s work describes Decision Intelligence as a bridge between decisions and data and develops the action-to-outcome framing.3 Pratt and Nadine Malcolm’s handbook brings causal models, technology and human knowledge together around complex decisions.4
Clayton Christensen’s Jobs to Be Done theory supports the narrower proposition that leaders should understand the progress people seek before designing a service response.5
Strasys interpretation
The definition used in this article is our healthcare-management definition of Decision Intelligence.
The proposition that Decision Intelligence reconstructs a whole decision from evidence held in separate functions is also our interpretation. So is the distinction between the roles of AI, expertise and accountable judgement.
SMASH and the Decision Continuum are our formal methods. Decision Intelligence as a field is not. It was named and developed outside Strasys, and we apply it across health and care systems internationally.
Limitations
Decision Intelligence cannot compensate for unreliable evidence, badly framed options or weak implementation. Models create false confidence when assumptions are hidden or uncertainty is understated. A well-structured decision can still be wrong.
The discipline should expose uncertainty, retain human accountability and test whether the chosen action produced the expected outcome.